About

About

Hi,

I’m Wingston, an engineering manager at Booking.com. I joined in 2017 as a software engineer and have led teams since 2018. Most of the job is hiring people and growing them into work they haven’t done before. I’ve hired 5 engineers, and 3 of my people have been promoted.

Since January 2025 I’ve run Flights Front End Target Architecture. It’s a platform team for developer experience and the front end. We own delivery metrics for more than 2,000 developers. I also work on our guidelines for building with AI agents.

Before that I set up the Flights team that shipped Fly Anywhere. It launched about 6 months after kickoff. In an A/B test it added about 200 bookings a day. Earlier, in Accommodations, I led the move from old Perl code to Java services. My team also cut abusive cancellations by about 40%.

The rest of my time goes on a problem I keep running into at work: a model gives you a confident answer, and you have no way to check whether it did the work or made it up.

That question turned into Agentosaurus, a build system where AI agents write most of the code for a climate site. I use it to learn how agents fail. The agents talk to each other through pull requests. They merged over a thousand in 2026. It also turned into psychohistory, three years of asking whether you can forecast crowds the way you forecast weather.

Before Booking.com: Industry Networks in Dubai, Premiere Digital in Bangalore and Los Angeles, and WingzTV, a company I started young, where I taught a lot of juniors to write software and made almost no money doing it.

Since 2022 I’ve chaired Taste Before You Waste, a food-waste charity in Amsterdam. Its books went from €20,334 below zero to a €106,573 reserve. I live-code music at algoraves. I got a Kinect v2 working on macOS, which Microsoft never shipped a driver for.

What I’m researching on this year

These are Coursera courses, each tied to something I’m building.

  • GPU Programming, a 4-course specialisation from Johns Hopkins, finished October 2026. CUDA from threads up to multi-GPU libraries. I use it for the LOGOS training runs on a single 3090.
  • The Fundamentals of RDMA Programming, from NVIDIA, October 2026. Moving data between GPUs on different machines. That’s the step after one 3090, if donated compute comes through.
  • AWS Cloud Solutions Architect, a 4-course certificate from AWS, September 2026. I haven’t taken the AWS exam. It feeds the infrastructure side of what I offer.
  • AI Governance (IAPP AIGP training), from Packt, August 2026. This is the training course. I haven’t taken the IAPP exam. It sits under the testing and governance work.
  • Project Management, a 6-course certificate from Google, July 2026.
CUDA coursework from the GPU Programming specialisation.

The full list with credential links is on LinkedIn.

Looking for my next role. Happy to talk about any of the above.

Get in touch · What I can help with